Assessing Reliability and Geometric Influence in Organization of Structural Brain Connectivity
Poster No:
1555
Submission Type:
Abstract Submission
Authors:
Sunghyoung Hong1, Seulki Yoo2, Bo-yong Park3, Seok-Jun Hong4
Institutions:
1Sungkyunkwan university, Suwon, Republic of Korea, 2Convergence Research Institute, Sungkyunkwan University, Suwon, Republic of Korea, 3Inha University, Incheon, Republic of Korea, 4IBS Center for Neuroscience Imaging Research, Suwon, Republic of Korea
First Author:
Co-Author(s):
Introduction:
Understanding the functional mechanism of human cognitive systems prerequisites a precise mapping of underlying physical brain networks. Structural connectivity (SC) based on diffusion tractography has been instrumental in this purpose, as it allows to grasp the configuration of whole-brain physical connections in a single matrix form. However, the prohibitively high dimensionality of SC matrices, together with inconsistent results depending on the tractography-related parameters (e.g. tractography method, parcellation, filtering strategy) have hindered the effort to use this technique in capturing the principles of structural brain organization[1]. Here, we sought to address these issues by applying a low-dimensional connectome mapping approach on SC[2,3], which provides a smooth transition of connectivity profiles across the whole brain (also known as "connectome gradient"). Moreover, inspired by a recent study demonstrating a significant role of gross cortical geometry on human brain function[4], we further tested the association between this SC gradient and cortical geometry to investigate a fundamental constraint in the manifestation of the structural brain organization.
Methods:
We analyzed structure and diffusion MRI of 86 young-adult subjects from the Q3 release of the Human Connectome Project (HCP)[5] (51 females; age range: 22-36). To generate SC matrices, we used three different tractography algorithms (FACT and SD STREAM: both for deterministic; iFOD2 for probabilistic)[6], three parcellations (Schaefer 400[7], Glasser 360[8], and sub-parcellation of Desikan-Killiany 400[3,9]), and two filtering methods (SIFT2[6] and COMMIT2[10]). Gradients were then generated from these SCs using Brainspace, specifically based on the diffusion map algorithm for non-linear dimensionality reduction with zero sparsity[3]. To compare this gradient with cortical geometry, we also calculated eigenmodes from a group-averaged mid-thickness cortical surface using the Laplace-Beltrami operator[4]. Next, gradients and SC were reconstructed by cumulatively adding geometric eigenmodes weighted by their corresponding coefficients.
Results:
We profiled the gradients of SC from 18 different combinations of tractography algorithms (3 tractography methods×3 parcellations×2 filtering algorithms). Among these, we found that the gradient showed high reproducibility, particularly in the 1st-4th components, between different algorithmic combinations (Fig1A). When we examined the variance explained by each principal component, however, it turned out that the first 3 components account for the majority of connectome data (~44% on average; Fig1B). These 3 principal components showed 1) the left to right, 2) the anterior to posterior, and 3) the sensory to transmodal axes (Fig1C). This suggests that even if the original SC matrices were seemingly complex, their underlying structures could be recapitulated by a parsimonious set of low-dimensional bases. Notably, as hypothesized, when we visualized the tractography results at each gradient, all of them seemed greatly affected by cortical geometric effects (Fig1D). Our post-hoc analysis confirmed this visual inspection that there are strikingly similar patterns between the eigenmodes of cortical surface and SC gradients (Fig2A,1C). Indeed, when we reconstructed the SC gradient based on cortical eigenmodes, >95% of accuracy was achieved with a minimal number of eigenmodes (Fig2B,2D). This strong geometric effect was also observed in the SC matrix (Fig2C,2D).
Conclusions:
Here, our study systematically investigated the principles of SC organization using a non-linear dimensionality reduction approach. Our main finding, i.e. a strong geometric effect on SC gradients, suggests that the topology of structural brain networks may be predominantly determined by the connectivity from the adjacent cortical areas (e.g., a short-range U fibers), a potential signature of evolutionarily shaped brain organization that prioritizes energy efficiency.
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural) 1
Diffusion MRI Modeling and Analysis 2
Image Registration and Computational Anatomy
Segmentation and Parcellation
Neuroanatomy, Physiology, Metabolism and Neurotransmission:
White Matter Anatomy, Fiber Pathways and Connectivity
Keywords:
Cortex
Data analysis
Modeling
MRI
STRUCTURAL MRI
Tractography
WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC
1|2Indicates the priority used for review
Provide references using author date format
2. Margulies, D. S., Ghosh, S. S., Goulas, A., Falkiewicz, M., Huntenburg, J. M., Langs, G., ... & Smallwood, J. (2016). Situating the default-mode network along a principal gradient of macroscale cortical organization. Proceedings of the National Academy of Sciences, 113(44), 12574-12579.
3. Vos de Wael, R., Benkarim, O., Paquola, C., Lariviere, S., Royer, J., Tavakol, S., ... & Bernhardt, B. C. (2020). BrainSpace: a toolbox for the analysis of macroscale gradients in neuroimaging and connectomics datasets. Communications biology, 3(1), 103.
4. Pang, J. C., Aquino, K. M., Oldehinkel, M., Robinson, P. A., Fulcher, B. D., Breakspear, M., & Fornito, A. (2023). Geometric constraints on human brain function. Nature, 1-9.
5. WU-Minn, H. C. P. (2017). 1200 subjects data release reference manual. URL https://www. humanconnectome. org, 565.
6. J.-D. Tournier, R. E. Smith, D. Raffelt, R. Tabbara, T. Dhollander, M. Pietsch, D. Christiaens, B. Jeurissen, C.-H. Yeh, and A. Connelly. MRtrix3: A fast, flexible and open software framework for medical image processing and visualisation. NeuroImage, 202 (2019), pp. 116–37.
7. Schaefer, A., Kong, R., Gordon, E. M., Laumann, T. O., Zuo, X. N., Holmes, A. J., ... & Yeo, B. T. (2018). Local-global parcellation of the human cerebral cortex from intrinsic functional connectivity MRI. Cerebral cortex, 28(9), 3095-3114.
8. Glasser, M. F., Coalson, T. S., Robinson, E. C., Hacker, C. D., Harwell, J., Yacoub, E., ... & Van Essen, D. C. (2016). A multi-modal parcellation of human cerebral cortex. Nature, 536(7615), 171-178.
9. Desikan RS, Ségonne F, Fischl B, et al. An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. Neuroimage 2006;31:968-980.
10. Schiavi, S., Ocampo-Pineda, M., Barakovic, M., Petit, L., Descoteaux, M., Thiran, J. P., & Daducci, A. (2020). A new method for accurate in vivo mapping of human brain connections using microstructural and anatomical information. Science advances, 6(31), eaba8245.